A promise is either enforced, or it's a liability.
Policy, process, tests and monitoring all help. Only architecture and cryptography make a promise true by construction.
In 8th grade I decided I wanted to be a software engineer someday. Eleven years of building iOS for billions of people later, at WhatsApp, Instagram and Google, I'm still a little amazed it worked out. :)
The part of the work I love most is the promises. When an app says your upload will finish or your photo stays private, something in the code has to make that true, and I like finding out what. These days I ask the same question about AI systems. I also run a small company, Swadharma Labs.
Eleven years of iOS, on both sides of the manager line.
Worked across messaging, privacy and reliability on iOS: link previews, Disappearing Messages, and Communities, where I was the engineering manager through global launch. I led a messaging reliability effort where fixing stability ended up moving product metrics, and most recently set technical direction for privacy features across a 20+ engineer org.
Managed Stories Formats, then Stories Creation & Visual Messaging. Built Reels sharing and led its reliability sprint after global launch. Then I chose to go back to the IC track.
Tech lead for suggested actions in Google Photos on iOS. Wrote shared iOS components that other teams built on: a Google profile component used in 10+ apps and a photo upload framework used in 4+. Mentored new Googlers and interns.
iOS and mobile architecture, MVP engineering requirements and team setup for startups; technical due diligence for investors; Staff+ story audits for senior engineers.
Mock technical interviews and coaching for software engineers.
Policy, process, tests and monitoring all help. Only architecture and cryptography make a promise true by construction.
Dead networks, killed apps, old OS versions, a billion devices. If a guarantee survives the client, it survives anywhere.
I want to build things that respect human systems instead of just extracting attention. That's where my writing on trust debt and cognitive safety comes from.
An open-source agent skill that reads a design doc or PR and shows which promises are actually enforced. Launching on GitHub soon.
Engineering systems, and the systems we build around ourselves: mobile and privacy architecture, AI-generated iOS code, trust debt, careers.
YouTube →Tech journeys, starting with interviewing: why strong engineers fail behavioral rounds, and how I'd prepare in 2026.
Staff+ Behavioral Interview Toolkit →Open-source prompts, evidence patterns and story-bank structure for preparing on your own.
you'll find me in a dance class, in a pool, or stuck on a killer sudoku (it fixed my doom-scrolling). I also cook Bihari dishes with whatever looks good at the farmers market that week (butternut squash chana dal is next on the list).
Copy whichever fits your talk, podcast or post.
Shalini Sah is a Staff+ engineer who has built iOS at billion-user scale for over a decade, at WhatsApp, Instagram and Google.
Shalini Sah is a Staff+ engineer who has spent more than a decade building iOS at billion-user scale. At WhatsApp, Shalini worked across messaging, privacy and reliability on iOS for 2B+ people, from Disappearing Messages to the global launch of Communities; at Instagram, Shalini managed Stories teams before returning to the IC track; at Google, Shalini wrote shared iOS components used across 10+ apps. Shalini now writes about how products keep their promises, from mobile architecture to AI systems, and runs Swadharma Labs, advising startups and investors on mobile and AI engineering.